The central theme of this Project is that the Medications Development effort for cocaine abuse would be aided by improved methods for early Phase II trials in which preliminary indications of efficacy are sought. To date, numerous medications have been judged as promising based on uncontrolled, open-label pilot trials, but subsequent replication attempts in large-scale double-blind, placebo-controlled trials have been disappointing. Reliance upon simple open-label pilot trials may have been misleading in choosing medications for further testing. This project will develop a model for early Phase II testing in which a candidate medication is tested in an integrated program involving both the human laboratory and small scale, controlled clinical trials. The clinical trial will incorporate placebo control via a multiple-baseline design, and laboratory cue-response as one of the dependent measures. Simultaneously, as part of a currently funded RO1 project (DA06234, Marian Fischman, PI) the same candidate medication will be tested in non-treatment-seeking cocaine users in a laboratory cocaine choice paradigm. Funding is requested in this application to add the cue-exposure procedure to the cocaine-choice paradigm. The principal outcome measures will be the same in both the clinic and the laboratory, namely cocaine use, cocaine craving, and subjective and physiologic response to cocaine cues. A series of candidates will be tested, amoxapine, risperidone, and ritanserin, which possess combinations of atypical neuroleptic and antidepressant properties. The following specific aims will be addressed for each candidate medication: Aim #1: To explore whether the candidate medication reduces cocaine use, cocaine craving, cocaine subjective effects and subjective and physiologic responses to cocaine cues in treatment-seeking outpatient cocaine abusers, and whether it has an acceptable side effect profile. Aim #2: To explore whether the candidate medication reduces cocaine use, craving and subjective effects, and subjective and physiologic responses to cocaine cues in non-treatment-seeking cocaine abusers during the laboratory cocaine-choice paradigm. Aim #3: To examine the degree of similarity between effects of the candidate medication in the preliminary clinical trial versus in the human laboratory cocaine-choice model, and to determine whether the candidate is appropriate for further Phase II or Phase III trials.